Unemployment

If the entire labor market is characterized by insider power, greater insider power reduces employment by raising the wage and causing firms to move up their labor demand curves. Thus in this case the insider-outsider distinction provides a candidate explanation of unemployment.

The more realistic case, however, is for there to be insider power only in part of the labor market, with the rest relatively competitive. But even in this case, insider power can increase average unemployment. When some sectors offer higher wages than others, workers have an incentive to try to obtain jobs in those sectors. New entrants to the labor market are therefore slower to accept jobs in the competitive sector, and workers who have been laid off from the high-wage sector accept longer spells of unemployment before they give up hope of returning to their old jobs.

This reasoning suggests that the contracting considerations investigated in Section 10.5 may also increase average unemployment. In the model analyzed there, the employment of the workers represented in the contracts is efficient. But we ignored the issues of whether such arrangements cover

See Problems 10.11 through 10.13 for examples of the effects of wage dispersion.

The critical assumption of the model is that the outsiders and insiders wages are linked. Without this link, the firm can hire outsiders at the prevailing economy-wide wage. With inelastic labor supply, that wage is low in recessions and high in booms, and so the marginal cost of labor to the firm is highly procyclical.

The insider-outsider literature has not made a definitive case that outsiders and insiders wages are linked. Gottfries argues that such a link arises from the facts that the firm must be given some freedom to discharge insiders who are incompetent or shirking and that an excessive gap between insiders and outsiders wages would give the firm an incentive to take advantage of this freedom. Blanchard and Summers (1986) argue that the insiders are reluctant to aUow the hiring of large numbers of outsiders at a low wage because they realize that, over time, such a policy would result in the outsiders controUing the bargaining process. But it is far from clear that tying insiders and outsiders wages is the best way of dealing with these problems. If the economy-wide wage is sometimes far below wi - c, tying the insiders and outsiders wages is very costly. The firms and the insiders might therefore be better off if they instead agreed to some limitation on the firms ability to hire outsiders, or if they charged new hires a fee (and let the fee vary with the gap between Wj and the economy-wide wage). Thus we can conclude only that ifa link between insiders and outsiders wages can be established, insider-outsider considerations have potentially important implications.

10.7 Hysteresis

One of the building blocks of the previous model is the assumption that the insiders are always employed. This assumption is likely to fail in some situations, however. Most importantly, if the insiders bargaining power is sufficiently great, they will set the wage high enough to risk some unemployment: if the insiders are fully employed with certainty, there is a benefit but not a cost to them of raising the wage further. In addition, unusually large negative shocks to labor demand are likely to lead to some unemployment among the insiders.

Variations in employment can give rise to dynamics in the number of insiders. Under many institutional arrangements, workers who become unemployed eventually lose a say in wage-setting; likewise, workers who are hired eventually gain a role in bargaining. Thus a fall in employment caused by a decline in labor demand is likely to reduce the number of insiders, and a rise in employment is likely to increase the number of insiders. These changes in the number of insiders then affect future wage-setting and employment.

These ideas are developed formally by Blanchard and Summers (1986). Blanchard and Summers focus on Europe in the 1980s, where, they argue, the conditions for these effects to be relevant were satisfied: workers had a great deal of power in wage-setting, there were large negative shocks, and the rules and institutions led to some extent to the disenfranchisement from the bargaining process of workers who lost their jobs .

2-See also Gregory (1986) and Blanchard and Summers (1987).

the entire economy, and of how workers come to be represented in such arrangements. If there are two sectors, one with explicit or implicit contracts and one with employment and wages largely determined competitively, and if workers fare better in the contract sector, then again they have an incentive to accept greater unemployment to increase then chances of obtaining these high-quality jobs.

There is relatively little evidence concerning how important these mechanisms are to actual unemployment. Summers (1986b) argues that such wait unemployment IS central to the determination of average unemployment. He presents evidence both across U.S. states and over time that general measures of wage dispersion and measures of wage differences between "high-quality" and "low-quality" jobs are strongly associated with differences in average unemployment rates. This is precisely what one would expect if workers efforts to obtain jobs paying more than the market-clearing wage are an important source of unemployment. Thus the limited evidence we have suggests these models may offer a promising route to understanding unemployment.

Shocks to labor demand are modeled by assuming that A is random, which implies that is random. Specifically, Ct is assumed to take the form

Cf = QOff, (10.62)

where C is a component of Cf that is known when workers set the wage and Et is an i.i.d. random shock that is determined after Wr is set.

In setting the wage, the insiders face a tradeoff between the expected fraction of the membership that is employed and the wage conditional on being employed. To see the consequences of endogenous changes in the number of insiders in the strongest possible form, assume that the insiders period-f objective function is the expected fraction of the insiders who are employed times utility conditional on being employed, and that this utilit\ takes the form wf (0 < b < 1). Since the insiders are assumed to be hired first and the number of insiders hired cannot exceed the number available, insider employment is the smaller of total employment and the number of

Assuming that insiders and outsiders wages differ by a constant, as in Section 10.6, has no important implications for the analysis.

Assumptions

We consider a simplified version of Blanchard and Summerss model. The wage is set unilaterally by the insiders, and employment is chosen by the firm. The number of insiders in one period is determined by the previous periods employment; thus

N,t = Lf-i. (10.58)

For simplicity, both the uisiders and the firm neglect the impact of their decisions on the future number of insiders; thus they maximize their current-period objective functions each period. The representative firms profits are

TTt = - Wt If, 0 < a < 1, (10.59)

where we assume for simplicity that all workers are paid the same wage, regardless of whether they are insiders. The first-order condition for the firms choice of employment is

aAtL}- = Wf. (10.60)

Solving (10.60) for I yields the labor demand curve:

l/(a-l)

(10.61)

= CfWt".